Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Can I recreate this polar coordinate spider chart in plotly?

Tags:

r

plotly

I'm having a bit of difficulty figuring out how to recreate the following graphic of a spider (or radar) chart, using plotly. Actually, I can't even recreate it in the most recent versions of ggplot2 because there have been breaking changes since 1.0.1.

Here's an example graphic:

enter image description here

Here's the original function that built it:

http://pcwww.liv.ac.uk/~william/Geodemographic%20Classifiability/func%20CreateRadialPlot.r

Here's an example of how the original function worked:

http://rstudio-pubs-static.s3.amazonaws.com/5795_e6e6411731bb4f1b9cc7eb49499c2082.html

Here's some not so dummy data:

d <- structure(list(Year = rep(c("2015","2016"),each=24),
                    Response = structure(rep(1L:24L,2), 
                                         .Label = c("Trustworthy", "Supportive", "Leading",
                                                    "Strong", "Dependable", "Consultative",
                                                    "Knowledgeable", "Sensible", 
                                                    "Intelligent", "Consistent", "Stable", 
                                                    "Innovative", "Aggressive", 
                                                    "Conservative", "Visionary", 
                                                    "Arrogant", "Professional", 
                                                    "Responsive", "Confident", "Accessible", 
                                                    "Timely", "Focused", "Niche", "None"),
                                         class = "factor"), 
                    Proportion = c(0.54, 0.48, 0.33, 0.35, 0.47, 0.3, 0.43, 0.29, 0.36,
                                   0.38, 0.45, 0.32, 0.27, 0.22, 0.26,0.95, 0.57, 0.42, 
                                   0.38, 0.5, 0.31, 0.31, 0.12, 0.88, 0.55, 0.55, 0.31,
                                   0.4, 0.5, 0.34, 0.53, 0.3, 0.41, 0.41, 0.46, 0.34, 
                                   0.22, 0.17, 0.28, 0.94, 0.62, 0.46, 0.41, 0.53, 0.34, 
                                   0.36, 0.1, 0.84), n = rep(c(240L,258L),each=24)),
               .Names = c("Year", "Response", "Proportion", "n"), 
               row.names = c(NA, -48L), class = c("tbl_df", "tbl", "data.frame"))

Here's my attempt (not very good)

plot_ly(d, r = Proportion, t = Response, x = Response, 
        color = factor(Year), mode = "markers") %>%
layout(margin = list(l=50,r=0,b=0,t=0,pad = 4), showlegend = TRUE)

Any thoughts on how I might be able to recreate this using plotly?

like image 759
Brandon Bertelsen Avatar asked Jun 07 '16 04:06

Brandon Bertelsen


People also ask

Is spider chart and radar chart same?

The radar chart is also known as web chart, spider chart, spider graph, spider web chart, star chart, star plot, cobweb chart, irregular polygon, polar chart, or Kiviat diagram. It is equivalent to a parallel coordinates plot, with the axes arranged radially.

What are spider charts good for?

A spider chart, also sometimes called a radar chart, is often used when you want to display data across several unique dimensions. Although there are exceptions, these dimensions are usually quantitative, and typically range from zero to a maximum value.

How do spider graphs work?

Spider diagrams are visual tools used to organize data in a logical way. A main concept is laid out on a page and lines are used to link ideas. As more ideas branch out, you're left with a graphical representation of something that may otherwise be difficult to understand.

What is a spider web graph called?

A Radar Chart, also called as Spider Chart, Radial Chart or Web Chart, is a graphical method of displaying multivariate data in the form of a two-dimensional chart of three or more quantitative variables represented on axes starting from the same point. ( Source: Wikipedia)


2 Answers

The options available with polar plots are still limited. There is not, so far as I can tell, any way to add text to a polar plot for the category labels around the circumference. Neither text scatter points, nor annotations nor tick labels (except at the four quarter points) are compatible with polar coordinates in plotly at the moment.

So, we need to get a little creative.

One type of polar coordinate system that does work nicely is a projected map of a sperical earth using an azimuthal projection. Here is a demonstration of how you might adapt that to this problem.

First, convert the values to plot into latitude and longitudes centred on the South pole:

scale <- 10   # multiply latitudes by a factor of 10 to scale plot to good size in initial view
d$lat <- scale*d$Proportion - 90
d$long <- (as.numeric(d$Response)-1) * 360/24

Plot using an azimuthal equidistant projection

p <- plot_ly(d[c(1:24,1,25:48,25),], lat=lat, lon=long, color = factor(Year), colors=c('#F8756B','#00BDC2'),
             type = 'scattergeo', mode = 'lines+markers', showlegend=T) %>%
layout(geo = list(scope='world', showland=F, showcoastlines=F, showframe=F,
             projection = list(type = 'azimuthal equidistant', rotation=list(lat=-90), scale=5)), 
             legend=list(x=0.7,y=0.85))

Put some labels on

p %<>% add_trace(type="scattergeo",  mode = "text", lat=rep(scale*1.1-90,24), lon=long, 
                 text=Response, showlegend=F, textfont=list(size=10)) %>%
       add_trace(type="scattergeo",  mode = "text", showlegend=F, textfont=list(size=12),
                 lat=seq(-90, -90+scale,length.out = 5), lon=rep(0,5), 
                 text=c("","25%","50%","75%","100%"))

Finally, add grid lines

l1 <- list(width = 0.5, color = rgb(.5,.5,.5), dash = "3px")
l2 <- list(width = 0.5, color = rgb(.5,.5,.5))
for (i in c(0.1, 0.25, 0.5, 0.75, 1)) 
    p <- add_trace(lat=rep(-90, 100)-scale*i, lon=seq(0,360, length.out=100), type='scattergeo', mode='lines', line=l1, showlegend=F, evaluate=T)
for (i in 1:24) 
    p <- add_trace(p,lat=c(-90+scale*0.1,-90+scale), lon=rep(i*360/24,2), type='scattergeo', mode='lines', line=l2, showlegend=F, evaluate=T)

enter image description here

Update for plotly version 4.x

Breaking changes in the updates to plotly mean that the original version no longer works without a few modifications to bring it up to date. here is an updated version:

library(data.table)
gridlines1 = data.table(lat = -90 + scale*(c(0.1, 0.25, 0.5, 0.75, 1)))
gridlines1 = gridlines1[, .(long = c(seq(0,360, length.out=100), NA)), by = lat]
gridlines1[is.na(long), lat := NA]

gridlines2 = data.table(long = seq(0,360, length.out=25)[-1])
gridlines2 = gridlines2[, .(lat = c(NA, -90, -90+scale, NA)), by = long]
gridlines2[is.na(lat), long := NA]

text.labels = data.table(
  lat=seq(-90, -90+scale,length.out = 5),
  long = 0,
  text=c("","25%","50%","75%","100%"))

p = plot_ly() %>%
add_trace(type="scattergeo", data = d[c(1:24, 1, 25:48, 25),], 
      lat=~lat, lon=~long, 
      color = factor(d[c(1:24, 1, 25:48, 25),]$Year), 
      mode = 'lines+markers')%>%
layout(geo = list(scope='world', showland=F, showcoastlines=F, showframe=F,
    projection = list(type = 'azimuthal equidistant', rotation=list(lat=-90), scale=5)), 
    legend = list(x=0.7, y=0.85)) %>%
add_trace(data = gridlines1, lat=~lat, lon=~long, 
    type='scattergeo', mode='lines', line=l1, 
    showlegend=F, inherit = F)  %>%
add_trace(data = gridlines2, lat=~lat, lon=~long,
    type='scattergeo', mode='lines', line=l2, showlegend=F) %>%
add_trace(data = text.labels, lat=~lat, lon=~long, 
  type="scattergeo", mode = "text", text=~text, textfont = list(size = 12),
    showlegend=F, inherit=F) %>%
add_trace(data = d, lat=-90+scale*1.2, lon=~long, 
    type="scattergeo", mode = "text", text=~Response, textfont = list(size = 10),
    showlegend=F, inherit=F) 

p
like image 121
dww Avatar answered Oct 18 '22 05:10

dww


I've made some progress with this, by faking it. Polar coords, seem to just hate me:

data:

df <- d <- structure(list(Year = c("2015", "2015", "2015", "2015", "2015", 
"2015", "2015", "2015", "2015", "2015", "2015", "2015", "2015", 
"2015", "2015", "2015", "2015", "2015", "2015", "2015", "2015", 
"2015", "2015", "2015", "2016", "2016", "2016", "2016", "2016", 
"2016", "2016", "2016", "2016", "2016", "2016", "2016", "2016", 
"2016", "2016", "2016", "2016", "2016", "2016", "2016", "2016", 
"2016", "2016", "2016"), Response = structure(c(1L, 2L, 3L, 4L, 
5L, 6L, 7L, 8L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L, 17L, 18L, 
19L, 20L, 21L, 22L, 23L, 24L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 
9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L, 17L, 18L, 19L, 20L, 21L, 
22L, 23L, 24L), .Label = c("Trustworthy", "Supportive", "Leading", 
"Strong", "Dependable", "Consultative", "Knowledgeable", "Sensible", 
"Intelligent", "Consistent", "Stable", "Innovative", "Aggressive", 
"Conservative", "Visionary", "Arrogant", "Professional", "Responsive", 
"Confident", "Accessible", "Timely", "Focused", "Niche", "None"
), class = "factor"), Proportion = c(0.54, 0.48, 0.33, 0.35, 
0.47, 0.3, 0.43, 0.29, 0.36, 0.38, 0.45, 0.32, 0.27, 0.22, 0.26, 
0.95, 0.57, 0.42, 0.38, 0.5, 0.31, 0.31, 0.12, 0.88, 0.55, 0.55, 
0.31, 0.4, 0.5, 0.34, 0.53, 0.3, 0.41, 0.41, 0.46, 0.34, 0.22, 
0.17, 0.28, 0.94, 0.62, 0.46, 0.41, 0.53, 0.34, 0.36, 0.1, 0.84
), n = c(240L, 240L, 240L, 240L, 240L, 240L, 240L, 240L, 240L, 
240L, 240L, 240L, 240L, 240L, 240L, 240L, 240L, 240L, 240L, 240L, 
240L, 240L, 240L, 240L, 258L, 258L, 258L, 258L, 258L, 258L, 258L, 
258L, 258L, 258L, 258L, 258L, 258L, 258L, 258L, 258L, 258L, 258L, 
258L, 258L, 258L, 258L, 258L, 258L)), .Names = c("Year", "Response", 
"Proportion", "n"), row.names = c(NA, -48L), class = c("tbl_df", 
"tbl", "data.frame"))

Create a circular mapping on a scatterplot, using basics:

df$degree <- seq(0,345,15) # 24 responses, equals 15 degrees per response
df$o <- df$Proportion * sin(df$degree * pi / 180) # SOH
df$a <- df$Proportion * cos(df$degree * pi / 180) # CAH
df$o100 <- 1 * sin(df$degree * pi / 180) # Outer ring x
df$a100 <- 1 * cos(df$degree * pi / 180) # Outer ring y 
df$a75 <- 0.75 * cos(df$degree * pi / 180) # 75% ring y
df$o75 <- 0.75 * sin(df$degree * pi / 180) # 75% ring x
df$o50 <- 0.5 * sin(df$degree * pi / 180) # 50% ring x
df$a50 <- 0.5 * cos(df$degree * pi / 180) # 50% ring y

And plot. I cheated here to get them to connect in the last position by double plotting row 1 and 25 again:

p = plot_ly()

for(i in 1:24) {
  p <- add_trace(
    p, 
    x = c(d$o100[i],0), 
    y = c(d$a100[i],0), 
    evaluate = TRUE,
    line = list(color = "#d3d3d3", dash = "3px"),
    showlegend = FALSE
    )
}

p %>% 
  add_trace(data = d[c(1:48,1,25),], x = o, y = a, color = Year, 
            mode = "lines+markers",
            hoverinfo = "text", 
            text = paste(Year, Response,round(Proportion * 100), "%")) %>% 
  add_trace(data = d, x = o100, y = a100, 
            text = Response,
            hoverinfo = "none",
            textposition = "top middle", mode = "lines+text", 
            line = list(color = "#d3d3d3", dash = "3px", shape = "spline"),
            showlegend = FALSE) %>% 
  add_trace(data = d, x = o50, y = a50, mode = "lines", 
            line = list(color = "#d3d3d3", dash = "3px", shape = "spline"), 
            hoverinfo = "none",
            showlegend = FALSE) %>% 
  add_trace(data = d, x = o75, y = a75, mode = "lines", 
            line = list(color = "#d3d3d3", dash = "3px", shape = "spline"), 
            hoverinfo = "none",
            showlegend = FALSE) %>%
  layout(
    autosize = FALSE,
    hovermode = "closest",     
    autoscale = TRUE,
    width = 800,
    height = 800,
    xaxis = list(range = c(-1.25,1.25), showticklabels = FALSE, zeroline = FALSE, showgrid = FALSE),
    yaxis = list(range = c(-1.25,1.25), showticklabels = FALSE, zeroline = FALSE, showgrid = FALSE))

As you can see, I've got it with the exception of that last connecting line, and the lines that pass from the origin to the text of the response.

enter image description here

like image 6
Brandon Bertelsen Avatar answered Oct 18 '22 04:10

Brandon Bertelsen